94 THE JOUKNAI, OF BOTANY 



I find it possible to identify the G. purpureum of Forster quite 

 satisfactorily, regarding Syme, not Forster, as responsible for any 

 confusion ; 1 also consider that the name var. rubricaule Horneni. 

 should not at present be used to replace var. maritimum. 



It is not necessary here to say much about the true G. purpureum 

 Vill. Villars originally included two plants under the name, one from 

 Buis, the other from the Pont de Claix near Grenoble. The former 

 was separated by Jordan in 1848 (in Adnot. Cat. pi. Jard. Grcnob. 3, 

 non vidi, ex Limnea, xxiii. 511 : 1850) as G. Villarsianum, a 

 Mediterranean form which does not occur in these islands. The 

 Pont de Claix plant was taken as the type of the emended G. pur- 

 pureum ; Jordan's account is to be found in .Bull. Soc' hot. Fr. vii. 

 005 (1800), in a note read before a meeting of the Society at 

 Grenoble, when Verlot exhibited dried specimens from the original 

 locality. In the National Herbarium are good series collected by 

 Verlot, labelled " G. purpureum Vill. rochers du pont de Claix, pres 

 de Grenoble," flowers collected in June 1807, fruits July 1800 ; with 

 other identical series from the Bastille of Grenoble collected in 1800 

 and 1802, the former labelled " Geranium purpureum Vill. — Jordan ! 

 Bulletin de la Societe bot. de France, session de Grenoble, 1800." In 

 the following account, Jordan's description and Verlot's specimens are 

 used as being the true G. purpureum Vill., emend Jord. 



Comparing G. purpureum with G. modestum Jord. (in Bull. Soc. 

 bot. Fr. vii. 005 : 1800), the only points of distinction which Jordan 

 himself found were " taille plus elevee, le port dresse, et . . . . les 

 petales . . . de forme oblongue et d'un pourpre vif " of G. modestum, 

 as contrasted with " port plus etale . . . fleurs rosees-purpurines " of 

 G. purpureum, the petal shape of the latter not being given. Kouy 

 says that the petals of G. purpureum {(jeuuinum) are also oblong. 

 The differences are therefore slight, and the flowers of all the 

 British specimens I have seen alive were more rose-purplish than 

 bright purple, while the plants were stiffly erect ; hence 1 agree with 

 Grenier, who long ago (in Mem. Soc. Emul. Doubs) united the two. 



Kouy, as I pointed out to Mr. Evans, made a mistake in referring 

 G. modestum to the group with the fruits covered with thick weals 

 instead of mere ridges. That group appears to be confined to the 

 Mediterranean region, and I regard it as forming a third distinct 

 species, G. Villarsianum Jord. emend., including G. semiylabrum 

 Jord. ex Boreau and G. mediterraneum Jord. 



G. Lebelii Boreau was distinguished from G. purpureum and 

 G. modestum on account of its hairy fruits. I find that specimens 

 of G. purpureum from the type-locality have the fruits sometimes 

 glabrous and sometimes hairy (even on the same shoot), and Rouy 

 states the same about modestum. The same is true of specimens from 

 most of the British localities — a fact which occasioned trouble to 

 critics in the Exchange Club Reports. 



I have seen G. purpureum Vill. from the following localities : — 

 Cornwall: Crantock, Newquay, Padstow, Wadebridge. Devonshire: 

 Torquay (the rock walk), ? Holes Hole, Bere Ferris (T. A. Briggs 

 collected some peculiar forms there, one of which, I think, may he 

 true G. purpurettm). Dorsetshire: Pirn field near Swanage (Kidley 



